Switching PSU for amplifiers

[Gold_xyz]

I would like to modify an old PC's supply for +/- 50V
or build a new PSU switching for my amplifier.

(today a cheap PC supply can deliver over 500W )

I look for to realize a light and compact amplifier but without abdicating the quality of class AB...

some suggestion?

[Gold_xyz]

Now I have taken around 10 PSU burnt by my retailer of computer.

I do have to resolve some fundamental problems...

- I need to rewind the trasformer for 50V
- Output diode schottky can to have insufficient Vmax
- I need to modify the optocouple feedback

[luka]

Hi

That should be fun.I have mod my to give me 13.5v, but it can easly be done for other voltages. What you need to do is to remember primary turns, and sec, wind it x that you have on, to get to 50. If there are 3 to get 5v I would try to put 30, but you should do some calculations, because probably you will need less than just multiplied x10.

Here great site to calculate for unregulated, you just add let say 20% more just to be sure.

[AndrewT]

Hi,
I have been trying to put others off the idea of SMPS for ClassAB amps. Part due to interaction between supply and active load and part to do with safety for the user/builder.

Can you guys show how it can be done safely?

I just opened up my mother's TV to find the supply was a direct online rectifier feeding a 400Vdc capacitor that must have been running at near maximum voltage to feed the SMPS.
No earth in sight and just a fuse blown. All working again.

Simple, light weight, cheap, all in no doubt. Help convert the biased.

[luka]

Hi

Let me get this right, you want to remake tv smps to operate at different output voltage, right?

[danville]

I have never considered a computer supply as a take off point, but I do think that a switching power supply for a linear amplifier is a good idea. It seems to me that this might be harder than it sounds because a computer supply is designed for high current at much lower voltages. The other catch is that it may be difficult to duplicate from one supply to the next.

With a conventional 50/60 Hz unregulated supply, you can only replenish the rails at a 100/120Hz rate. The power supply caps are never really big enough.

If the switcher operates above the audio frequency range, the supply will never sag very much. You still need to have good power supply rejection in the VAS stage. This will be a little more difficult since the power supply noise will be at a higher frequency, but I think the amplifier will probably perform better.

I've built lots of switching supplies but nothing higher power. I think this might be an interesting area to explore further.

From a safety point of view, I don't think this is much different than any other type of supply. Don't touch the rail voltages......

Perhaps this topic has discussed at length in the Class D section.

Al Clark

[DSP_Geek]

Originally posted by Gold_xyz
I would like to modify an old PC's supply for +/- 50V
or build a new PSU switching for my amplifier.

(today a cheap PC supply can deliver over 500W )

I look for to realize a light and compact amplifier but without abdicating the quality of class AB...

some suggestion?

I know my way around a soldering iron, yet the books I picked up on switching power supply design convinced me I didn't want to play around with something which could fry me stone dead in a second. There's a reason SMPS consultants get $150/hour.

On the other hand, if you really want a switcher, try these guys:

http://www.coldamp.com/opencms/openc.../en/productos/

[Conrad Hoffman]

Gee, I've built switchers at up over 1kV, but never got $150/hour for it. Bummer. IMO, you probably want to design this from scratch, rather than fooling with a PC supply. Aside from poor docs, you'll probably have to replace every part in it. A better place to start is the switchers used in any high power automotive amp. Use the basic concept, but redo the transformer design for the right ratio for line in, and be sure the feedback system is isolated like optocouplers or such. With small loop areas and a ground plane, it should be possible to make a very usable supply. Unless you're really good (I'm not), don't expect the fantastic efficiency number you sometimes see advertised.

[DSP_Geek]

Originally posted by Conrad Hoffman
Gee, I've built switchers at up over 1kV, but never got $150/hour for it. Bummer. IMO, you probably want to design this from scratch, rather than fooling with a PC supply. Aside from poor docs, you'll probably have to replace every part in it. A better place to start is the switchers used in any high power automotive amp. Use the basic concept, but redo the transformer design for the right ratio for line in, and be sure the feedback system is isolated like optocouplers or such.

There's also the small matter of replacing all the primary-side silicon because car power supplies are 12 volts while rectified wall voltage is about 170 Vdc.

[Pierre]

Well, Conrad, that's not what I call a friend's advice!

Toroids used in car SMPS can be used because there is no HV involved. You won't be able to keep reasonable safety with a toroid used in an offline PSU. I wouldn't play with that, and start with a completely new ETD core instead.
All the primary side has to be changed, as DSP_geek noticed (not only the components voltage ratings, but also the topology, car SMPSs are usually push-pull, while you want a half-bridge or full-bridge for offline supply).

In a nutshell, the only part you can reuse of a car amp is the secondary side, excluding the transformer.